agentleFS
vfarcicvfarcic/dot-ai/.claude/skills/write-docs/SKILL.md

write-docs

Skill336 starsChanged 10 months agoMITView on GitHub

Write documentation with real, validated examples. Executes commands through the user to capture actual output. Use for any new documentation or major doc updates.

45 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use close-spec, design-with-images and 9 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • dzhngSkill

    close-spec

    When a spec under specs/<feature>/ is done shipping (a write-spec build, or any planned task that produced a spec), archive it to specs/done/ and rewrite it from a build-plan into a durable rationale — the why, the principles, the invariants — pointing back to the real code for the how. Use when implementation has landed and the plan no longer matches what shipped, or the user says a feature/spec is finished. Pairs with [write-spec](../write-spec/SKILL.md) (the plan this closes) and [review](../review/SKILL.md) (run before closing).

    Also in 5 projects that use this

  • dzhngSkill

    design-with-images

    Explore UI and visual design with image generation, then iterate the real implementation against the user-selected concept. Use when brainstorming visual directions, offering mockup variants, or implementing an approved generated design; do not stop at the mockups or the first code approximation.

    Also in 5 projects that use this

  • transitive-bullshitSkill

    compare-screenshots

    Compare screenshots against the intended design, distinguishing approved references from historical baselines. Use for iterative implementation matching, before/after visual review, objective image telemetry, or checking a lone capture for flat, empty, or badly framed content.

    Also in 5 projects that use this

  • LucasMatuszewskiSkill

    audit-agents

    Audit and maintain agent configuration files (.claude/agents, AGENTS.md, CLAUDE.md, agent-memory). Checks for duplicates, outdated info, naming conflicts, wrong technical facts, and files over 200 lines. Fixes all issues and commits.

    Also in 5 projects that use this

  • wolverin0Skill

    audit-decisions

    Standardize remediation decisions, baseline-policy entries, and legal-review banners for accepted, deferred, or policy-blocked audit findings.

    Also in 4 projects that use this

  • rhnfzlSkill

    explore-unknowns

    Walk the user through mapping the unknowns of a task before, during, and after they build it, and hand them a four-quadrant map (settled ground, open decisions, taste, and landmines) they keep. Reach for this eagerly whenever a request is ambiguous or underspecified, the codebase or domain is unfamiliar, the user will "know it when they see it", a reference implementation must be understood before porting, a build-queue or architecture-changing task is about to start, a mid-build deviation needs capturing, or a finished change needs buy-in or verified understanding before merge. Trigger even when the skill is not named: "do a blindspot pass", "find my unknown unknowns", "interview me about this change", "brainstorm the approach for a task I am about to build", "help me scope this", "what am I missing here", "plan this before I build", "I know nothing about X in this codebase". When a territory profile is present it reads memory and the graph first and feeds the map into the build pipeline; without one it degrades to a plain general map. Not for trivial or mechanical tasks with clear acceptance criteria, single-fact lookups, pure debugging of a known symptom, or pure code review, those have their own tools. Do not use it to run the build itself; the map is the deliverable and implementation remains a separate task, whether it has not started or is already underway.

    Also in 5 projects that use this

  • +5 more from dzhng

The file

Show its rules and text

We split this skill into the rules it gives an agent

22 rules. 0 are common, 19 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 43 edited versions we read keep these.

  • 19

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write accurate, user-focused documentation with real examples by executing operations through the user.

    line 8
  2. Distinctive:

    Execute-then-document: Never write examples without real output

    line 12
  3. Distinctive:

    Chunk-by-chunk: Write one section at a time, get confirmation before proceeding

    line 13
  4. Distinctive:

    User-focused: Write for users, not developers

    line 14
  5. Distinctive:

    Validate prerequisites: Run setup steps from existing docs to verify they work

    line 15
  6. Distinctive:

    Fix docs when broken: If existing docs don't work during setup, STOP and discuss updating those docs before proceeding.

    line 17
  7. Distinctive:

    Ask the user what documentation to write.

    line 23
  8. Distinctive:

    If documenting a feature not yet in published charts:

    line 65
  9. Distinctive:

    Add these flags to helm install:

    line 77
  10. Very specific:

    IMPORTANT: Always use KUBECONFIG=./kubeconfig.yaml for all kubectl/helm commands.

    line 87
  11. Distinctive:

    If any step fails or doesn't match existing docs: STOP and discuss whether to update those docs before proceeding.

    line 89
  12. Distinctive:

    MCP interactions: Ask user to send intent and share output (these are the user-facing examples we're documenting)

    line 116
  13. Distinctive:

    After all sections are written:

    line 132
  14. Distinctive:

    Verify internal links work

    line 133
  15. Distinctive:

    Check links to other docs exist

    line 134
  16. Very specific:

    Update mcp-tools-overview.md if adding a new tool guide

    line 135
  17. Distinctive:

    Update any index pages

    line 136
  18. Distinctive:

    Never invent output: - Always use real responses from the user

    line 169
  19. Distinctive:

    Never skip validation: - Every example must be executed first

    line 170
  20. Distinctive:

    Use user language: - Avoid internal/developer terminology

    line 172
  21. Distinctive:

    Include error cases: - Document what happens when things go wrong

    line 173
  22. Very specific:

    Full flags in commands: - Always use long-form flags (e.g., --filename not -f, --namespace not -n, --output not -o).

    line 174

“Common” is measured: kept by at least 6 in 10 of the 43 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Write Documentation

Write accurate, user-focused documentation with real examples by executing operations through the user.

Principles

  1. Execute-then-document: Never write examples without real output
  2. Chunk-by-chunk: Write one section at a time, get confirmation before proceeding
  3. User-focused: Write for users, not developers
  4. Validate prerequisites: Run setup steps from existing docs to verify they work
  5. Claude runs infrastructure, user runs MCP: Claude executes all bash/infrastructure commands. Only ask user for MCP client interactions (the actual user-facing examples).
  6. Fix docs when broken: If existing docs don't work during setup, STOP and discuss updating those docs before proceeding.

Workflow

Step 1: Identify the Documentation Target

Ask the user what documentation to write. Options:

  • New feature guide (e.g., "knowledge base guide")
  • Update existing guide
  • API reference
  • Setup/configuration guide

Step 2: Fresh Environment Setup

ALWAYS start with a clean test cluster to ensure reproducible documentation.

Follow the actual docs (docs/setup/mcp-setup.md) - this validates they work.

Claude executes all infrastructure steps directly using Bash tool:

  1. Tear down existing test cluster if present
    kind delete cluster --name dot-ai-test 2>/dev/null || true
    rm -f ./kubeconfig.yaml
  2. Create fresh Kind cluster with local kubeconfig
    kind create cluster --name dot-ai-test --kubeconfig ./kubeconfig.yaml
    export KUBECONFIG=./kubeconfig.yaml
  3. Install prerequisites (ingress controller)
    kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/main/deploy/static/provider/kind/deploy.yaml
    # Wait for ingress to be ready
    kubectl wait --namespace ingress-nginx --for=condition=ready pod --selector=app.kubernetes.io/component=controller --timeout=300s
  4. Follow docs/setup/mcp-setup.md (skip controller if not needed for the feature)
    • Step 1: Set environment variables (use existing API keys from env)
    • Step 2: Install controller via Helm (skip if feature doesn't need it)
    • Step 3: Install MCP server via Helm
    • Step 4: Tell user to configure MCP client
    • Step 5: Tell user to verify with "Show dot-ai status"
  5. For unreleased features: Build and use local images

    If documenting a feature not yet in published charts:

    # Build MCP server image
    npm run build
    docker build -t dot-ai:test .
    kind load docker-image dot-ai:test --name dot-ai-test
    
    # Build agentic-tools plugin image
    docker build -t dot-ai-agentic-tools:test ./packages/agentic-tools
    kind load docker-image dot-ai-agentic-tools:test --name dot-ai-test

    Add these flags to helm install:

    --set image.repository=dot-ai \
    --set image.tag=test \
    --set image.pullPolicy=Never \
    --set plugins.agentic-tools.image.repository=dot-ai-agentic-tools \
    --set plugins.agentic-tools.image.tag=test \
    --set plugins.agentic-tools.image.pullPolicy=Never

IMPORTANT: Always use KUBECONFIG=./kubeconfig.yaml for all kubectl/helm commands.

If any step fails or doesn't match existing docs: STOP and discuss whether to update those docs before proceeding.

Why fresh cluster? Ensures documentation examples work from a known clean state and validates setup docs.

Step 3: Outline the Documentation

Present an outline of sections to write. Example:

1. Overview (what it does, when to use it)
2. Prerequisites
3. Basic Usage (with real examples)
4. Advanced Features
5. API Reference
6. Troubleshooting
7. Next Steps

Get user confirmation on the outline before proceeding.

Step 4: Write Chunk-by-Chunk

🚨 CRITICAL: One section at a time. NEVER write multiple sections or the whole doc at once.

For each section:

  1. Execute first — Run every command or ask user to run every MCP interaction that will appear in this section. Do this BEFORE writing any documentation prose.
    • Bash commands: Claude runs these directly using Bash tool
    • MCP interactions: Ask user to send intent and share output (these are the user-facing examples we're documenting)
  2. Wait for output — For MCP interactions, user shares the actual output. For bash commands, capture the real output.
  3. Write and apply the chunk immediately — Write the documentation section using real output and apply the edit directly (Write or Edit tool). Do NOT ask for permission — the user will cancel the edit if they want changes. Do NOT show the markdown in a code block and ask "should I write this?" — just write it.
  4. Proceed to next section — Move to the next section and repeat.

Key distinction: Infrastructure/setup = Claude runs it. User-facing MCP examples = User runs it and shares output.

NEVER do these:

  • ❌ Write the entire document in one go
  • ❌ Write multiple sections before getting confirmation
  • ❌ Show proposed markdown and ask "want me to write this?"
  • ❌ Write examples without executing the commands first
  • ❌ Skip execution for "simple" or "standard" commands — validate everything

Step 5: Cross-Reference Check

After all sections are written:

  • Verify internal links work
  • Check links to other docs exist
  • Update mcp-tools-overview.md if adding a new tool guide
  • Update any index pages

Step 6: Final Review

Tell the user: "Documentation complete. Please review the full file and let me know if any adjustments are needed."

Example Execution Request Formats

For MCP tool operations:

Please send this intent to your MCP client:
"Ingest this document into the knowledge base: [content] with URI: [url]"

Share the response you receive.

For status checks:

Please ask: "Show dot-ai status"

Share what you see for the Vector DB collections.

For bash commands:

Please run:
kubectl get pods -n dot-ai

Share the output.

Important Rules

  • Never invent output - Always use real responses from the user
  • Never skip validation - Every example must be executed first
  • Keep sections small - One concept per chunk for easy review
  • Use user language - Avoid internal/developer terminology
  • Include error cases - Document what happens when things go wrong
  • Full flags in commands - Always use long-form flags (e.g., --filename not -f, --namespace not -n, --output not -o). Full flags are more self-documenting for users unfamiliar with the tools.

File Locations

  • Feature guides: docs/guides/mcp-*-guide.md
  • Setup guides: docs/setup/*.md
  • Tool overview: docs/guides/mcp-tools-overview.md
  • Images: docs/img/

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.